carlosdanielhernandezmena's picture
Fixing widget error and bibtex format
3ca9037
metadata
language: is
datasets:
  - language-and-voice-lab/samromur_milljon
tags:
  - audio
  - automatic-speech-recognition
  - icelandic
  - xlrs-53-icelandic
  - iceland
  - reykjavik
  - samromur
license: cc-by-4.0
model-index:
  - name: wav2vec2-large-xlsr-53-icelandic-ep30-967h
    results:
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: Samrómur (Test)
          type: language-and-voice-lab/samromur_asr
          split: test
          args:
            language: is
        metrics:
          - name: WER
            type: wer
            value: 7.698
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: Samrómur (Dev)
          type: language-and-voice-lab/samromur_asr
          split: validation
          args:
            language: is
        metrics:
          - name: WER
            type: wer
            value: 6.786
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: Samrómur Children (Test)
          type: language-and-voice-lab/samromur_children
          split: test
          args:
            language: is
        metrics:
          - name: WER
            type: wer
            value: 6.467
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: Samrómur Children (Dev)
          type: language-and-voice-lab/samromur_children
          split: validation
          args:
            language: is
        metrics:
          - name: WER
            type: wer
            value: 4.234
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: Malrómur (Test)
          type: language-and-voice-lab/malromur_asr
          split: test
          args:
            language: is
        metrics:
          - name: WER
            type: wer
            value: 6.631
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: Malrómur (Dev)
          type: language-and-voice-lab/malromur_asr
          split: validation
          args:
            language: is
        metrics:
          - name: WER
            type: wer
            value: 5.836
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: Althingi (Test)
          type: language-and-voice-lab/althingi_asr
          split: test
          args:
            language: is
        metrics:
          - name: WER
            type: wer
            value: 17.904
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: Althingi (Dev)
          type: language-and-voice-lab/althingi_asr
          split: validation
          args:
            language: is
        metrics:
          - name: WER
            type: wer
            value: 17.931

wav2vec2-large-xlsr-53-icelandic-ep30-967h

The "wav2vec2-large-xlsr-53-icelandic-ep30-967h" is an acoustic model suitable for Automatic Speech Recognition in Icelandic. It is the result of fine-tuning the model facebook/wav2vec2-large-xlsr-53 for 30 epochs with 967 hours of Icelandic data collected by the Language and Voice Laboratory through the platform Samrómur.

The specific data that was used to fine-tune the model is the corpus Samrómur Milljón, which is the result of the automatic verification of 1 million of recordings comming from the corpus "Samromur Unverified 22.07". It has to be pointed out that this model was trained with different data than our previous model wav2vec2-large-xlsr-53-icelandic-ep10-1000h .

The fine-tuning process was performed during July (2023) in the servers of the Language and Voice Laboratory (https://lvl.ru.is/) at Reykjavík University (Iceland) by Carlos Daniel Hernández Mena.

Evaluation

import torch
from transformers import Wav2Vec2Processor
from transformers import Wav2Vec2ForCTC

#Load the processor and model.
MODEL_NAME="language-and-voice-lab/wav2vec2-large-xlsr-53-icelandic-ep30-967h"
processor = Wav2Vec2Processor.from_pretrained(MODEL_NAME)
model = Wav2Vec2ForCTC.from_pretrained(MODEL_NAME)

#Load the dataset
from datasets import load_dataset, load_metric, Audio
ds=load_dataset("language-and-voice-lab/samromur_children", split="test")

#Downsample to 16kHz
ds = ds.cast_column("audio", Audio(sampling_rate=16_000))

#Process the dataset
def prepare_dataset(batch):
    audio = batch["audio"]
    #Batched output is "un-batched" to ensure mapping is correct
    batch["input_values"] = processor(audio["array"], sampling_rate=audio["sampling_rate"]).input_values[0]
    with processor.as_target_processor():
        batch["labels"] = processor(batch["normalized_text"]).input_ids
    return batch
ds = ds.map(prepare_dataset, remove_columns=ds.column_names,num_proc=1)

#Define the evaluation metric
import numpy as np
wer_metric = load_metric("wer")
def compute_metrics(pred):
    pred_logits = pred.predictions
    pred_ids = np.argmax(pred_logits, axis=-1)
    pred.label_ids[pred.label_ids == -100] = processor.tokenizer.pad_token_id
    pred_str = processor.batch_decode(pred_ids)
    #We do not want to group tokens when computing the metrics
    label_str = processor.batch_decode(pred.label_ids, group_tokens=False)
    wer = wer_metric.compute(predictions=pred_str, references=label_str)
    return {"wer": wer}

#Do the evaluation (with batch_size=1)
model = model.to(torch.device("cuda"))
def map_to_result(batch):
    with torch.no_grad():
        input_values = torch.tensor(batch["input_values"], device="cuda").unsqueeze(0)
        logits = model(input_values).logits
    pred_ids = torch.argmax(logits, dim=-1)
    batch["pred_str"] = processor.batch_decode(pred_ids)[0]
    batch["sentence"] = processor.decode(batch["labels"], group_tokens=False)
    return batch
results = ds.map(map_to_result,remove_columns=ds.column_names)

#Compute the overall WER now.
print("Test WER: {:.3f}".format(wer_metric.compute(predictions=results["pred_str"], references=results["sentence"])))

Test Result: 0.076

BibTeX entry and citation info

When publishing results based on these models please refer to:

@misc{mena2023xlrs53icelandic30ep967h,
      title={Acoustic Model in Icelandic: wav2vec2-large-xlsr-53-icelandic-ep30-967h.}, 
      author={Hernandez Mena, Carlos Daniel},
      url={https://huggingface.co/language-and-voice-lab/wav2vec2-large-xlsr-53-icelandic-ep30-967h},
      year={2023}
}

Acknowledgements

Thanks to Jón Guðnason, head of the Language and Voice Lab for providing computational power to make this model possible.

We also want to thank to the "Language Technology Programme for Icelandic 2019-2023" which is managed and coordinated by Almannarómur, and it is funded by the Icelandic Ministry of Education, Science and Culture. This model is an unexpected result of all the resources gathered by the Programme.

Special thanks to Björn Ingi Stefánsson for setting up the configuration of the server where this model was trained.